Noise

[mchaos]

Hello NI mambers,

I am racking my brain about this one. I am using a SCCXI1303 terminal block connected to a SCXI1000 chassis and interfaced with a PCI E series board. I am trying to read the voltage signal from the following circuit.

Essentially the "probe" is an electrode with an open air gap. When the probe is inserted in an electrically conducting media (i.e. a plasma) an ion current is generated closing the circuit and generating a voltage across the 500 kOhm resistor. Current levels are small, on the order of 3 microamps. I am getting LOTS of noise here. Doing an FFT gives me a main 60 hZ component so I am thinking ground loop. Right now I am grounding at the 1303 chassis. I have tried other groundings to no avail. All components (power supply, DAQ, and DAQ computer) are connected to the same power supply. Any ideas?

Thanks in advance for your help.

Regards,

Marcos

[mchaos]

Forgot to mention that the noise is present even when the circuit is open.

[mchaos]

Also, the power supply provides 200VDC. Sorry about the missing details and thanks again.

[johnsold]

What are the pull up and bias resistors in your SCXI-1303?

How long are the wires to the probe? How is the plasma generated? Is the 200 VDC supply "clean"?  How much is LOTS of noise?  Quantitative measurements are much more helpful than qualitative descriptions when it comes to figuring out what is going on.

I suggest that you also consider adding some protection circuits.  If the probe ever shorts, you will have 200 V at the input to your SCXI-1303, which will almost certainly destroy it!

Please post your VI with some typical data saved as default.

Lynn

[mchaos]

Thank you for your response Lynn,

For shakedown testing purposes I am using a flame from a lighter to generate the "plasma." In fact the system will be used to detect flames. The lead wires to the probe are about 7 ft long, no way arouind that given the system setup. The 200VDC supply can indeed be considered "clean." The pullup and bias resistors in the 1303 are 10 MOhms; these can be changed or removed altogether.

I'll give you an example of the "lots" of noise I refer to. With an open circuit (i.e. no flame) the system reads an oscillatory signal with an average of about 0.2V and an apparent amplitude of 0.3V. As I posted, the FFT of the signal has a main 60Hz component (along with its overtones). I do not have access to the data at the moment; I can post a plot tomorrow when I get to the lab. The VI I use is quite big with lots of sub VIs so I do not know if you want to go through it but I can definetely send it to you as a zip file if needed. I hear your concern about shorting the probe; I just want to take care of this issue and then consider a protection circuit.

Thank you again for your help.

Regards,

Marcos

[johnsold]

Marcos,

In the open condition you should have about 238 mV (DC) because of the 10 megohm pull up resistor. The bias resistor is shorted by grounding the probe cable.  The pull up resistor will also result in about 165 mV of error when the probe is conducting 3 uA.

With the probe open no current is flowing so any noise pick up is due to electric fields rather than magnetic fields.  Try twisting the wires to the probe tightly together.  Twisting works better for magnetic coupling but will not hurt.  Next try shielding the wires.  Connect the shield to ground only at the point where the probe wire is grounded at the 1303.  Shielding can be quite effective against electric fields.  Consider the exposure of the wire from the 500 kilohm resistor to the power supply also.  That should be shielded, too.

Lynn

[mchaos]

Lynn,

Excellent suggestion. As it stands now, the circuit was built haphazardly so I will definitely take care in proper shielding and grounding. Thanks again for your help.

Marcos

[Henrik_Volkers]

Shortcut the input with two antiparallel diodes (1N4148,,1n4007) to limit the input voltage to 700mV and add a 1k resisitor for current limiting from the not grounded input to the voltage source

If the voltagesource is to noisy you can try RC filtering you DC supply with 10k and 220µV/260V (and add a 470k bleeding resistor over the cap 😉 ) this slightly reduce the 200V but reduce voltage ripple. More on low noise high voltage supplies in linear.com appnote AN-118 ..